The financial implications of cardiac stress testing prior to abdominal aortic aneurysm repair.

BACKGROUND: The utilization and cost-effectiveness of stress testing before abdominal aortic aneurysm (AAA) repair remains insufficiently studied. We examined the variation and financial implications of stress testing, and their association with major adverse cardiovascular events (MACE). METHODS: We studied patients who underwent elective endovascular (EVAR) or open AAA repair (OAR) at Vascular Quality Initiative centers from 2015 to 2019. We grouped centers into quintiles of preoperative stress testing frequency. We calculated the risk of postoperative MACE, a composite of in-hospital myocardial infarction, heart failure, or death, for each center-quintile. We obtained charges for stress tests locally and applied these to the cohort to estimate charges per 1000 patients. RESULTS: We studied 32,459 patients (EVAR: 27,978; OAR: 4481; 283 centers). Stress test utilization varied across quintiles from 13.0% to 68.6% (median: 36.8%) before EVAR and 15.9% to 85.0% (median: 59.4%) before OAR. The risk of MACE was 1.4% after EVAR and 10.2% after OAR. There was a trend towards more common MACE after EVAR among centers with higher utilization of stress testing: 0.9% among centers in the lowest quintile, versus 1.7% in the highest quintile (p-trend = 0.068). There was no association between MACE and stress testing frequency for OAR (p-trend = 0.223). The estimated financial charges for stress testing before EVAR ranged from $125,806 per 1000 patients at 1st-quintile centers, to $665,975 at 5th-quintile centers. Charges before OAR ranged from $153,861 at 1st-quintile centers, to $825,473 at 5th-quintile centers. CONCLUSION: Stress test use before AAA repair is highly variable and associated with substantial cost, with an unclear association with postoperative MACE. This highlights the need for improved stress testing paradigms prior to surgery.

Use of a support vector machine for categorizing free-text notes: assessment of accuracy across two institutions.

BACKGROUND: Electronic health record (EHR) users must regularly review large amounts of data in order to make informed clinical decisions, and such review is time-consuming and often overwhelming. Technologies like automated summarization tools, EHR search engines and natural language processing have been shown to help clinicians manage this information. OBJECTIVE: To develop a support vector machine (SVM)-based system for identifying EHR progress notes pertaining to diabetes, and to validate it at two institutions. MATERIALS AND METHODS: We retrieved 2000 EHR progress notes from patients with diabetes at the Brigham and Women's Hospital (1000 for training and 1000 for testing) and another 1000 notes from the University of Texas Physicians (for validation). We manually annotated all notes and trained a SVM using a bag of words approach. We then used the SVM on the testing and validation sets and evaluated its performance with the area under the curve (AUC) and F statistics. RESULTS: The model accurately identified diabetes-related notes in both the Brigham and Women's Hospital testing set (AUC=0.956, F=0.934) and the external University of Texas Faculty Physicians validation set (AUC=0.947, F=0.935). DISCUSSION: Overall, the model we developed was quite accurate. Furthermore, it generalized, without loss of accuracy, to another institution with a different EHR and a distinct patient and provider population. CONCLUSIONS: It is possible to use a SVM-based classifier to identify EHR progress notes pertaining to diabetes, and the model generalizes well.